Data Platform Engineer

Join a new, small team building PerfectServe’s data platform. You will work closely with a staff data engineer and a platform engineer, with enough ownership to see your work in production and enough support to learn quickly.

The source systems are varied, so the work spans extending established pipelines and designing new ones from scratch. That range is a good part of the job.

Why this role, specifically

Joining a platform in its first months is a different experience from joining one that is five years old. The patterns are still being set, which means your questions change them. You will see the whole path — a production source system at one end, someone making a decision at the other — rather than one layer of a stack somebody else assembled.

You will have a dedicated team to work with and learn from so the ownership comes without being thrown in unsupported.

What you will do

• Build and maintain ingestion pipelines from production source systems
• Write transformation models that turn raw data into datasets people can trust
• Apply dimensional and medallion modelling patterns as the warehouse takes shape
• Deploy pipelines to production, and test them properly — idempotency, data quality checks
• Contribute to data model and schema design, and to design reviews
• Document what you build, including runbooks for the pipelines you own

What you bring

• Professional experience building data pipelines, with production ownership of at least some
• Strong SQL, and a working programming language — Python, Go or equivalent
• Hands-on dbt, or another transformation framework
• A cloud data warehouse — Snowflake, BigQuery, Redshift or similar
• Practical ELT and orchestration experience (Fivetran, Airflow or equivalent)
• Working AWS knowledge — object storage, access basics, and enough networking to reason about where data moves
• A testing instinct for data, not only for code

Nice to have

• Change-data-capture
• SQL Server sources
• Healthcare or another regulated data environment
• Exposure to a BI or semantic layer